2 bup: It backs things up
3 =======================
5 bup is a program that backs things up. It's short for "backup." Can you
6 believe that nobody else has named an open source program "bup" after all
9 Despite its unassuming name, bup is pretty cool. To give you an idea of
10 just how cool it is, I wrote you this poem:
13 What rhymes with awesome?
15 But that's irrelevant.
17 Hmm. Did that help? Maybe prose is more useful after all.
20 Reasons bup is awesome
21 ----------------------
23 bup has a few advantages over other backup software:
25 - It uses a rolling checksum algorithm (similar to rsync) to split large
26 files into chunks. The most useful result of this is you can backup huge
27 virtual machine (VM) disk images, databases, and XML files incrementally,
28 even though they're typically all in one huge file, and not use tons of
29 disk space for multiple versions.
31 - It uses the packfile format from git (the open source version control
32 system), so you can access the stored data even if you don't like bup's
35 - Unlike git, it writes packfiles *directly* (instead of having a separate
36 garbage collection / repacking stage) so it's fast even with gratuitously
37 huge amounts of data. bup's improved index formats also allow you to
38 track far more filenames than git (millions) and keep track of far more
39 objects (hundreds or thousands of gigabytes).
41 - Data is "automagically" shared between incremental backups without having
42 to know which backup is based on which other one - even if the backups
43 are made from two different computers that don't even know about each
44 other. You just tell bup to back stuff up, and it saves only the minimum
45 amount of data needed.
47 - You can back up directly to a remote bup server, without needing tons of
48 temporary disk space on the computer being backed up. And if your backup
49 is interrupted halfway through, the next run will pick up where you left
50 off. And it's easy to set up a bup server: just install bup on any
51 machine where you have ssh access.
53 - Bup can use "par2" redundancy to recover corrupted backups even if your
54 disk has undetected bad sectors.
56 - Even when a backup is incremental, you don't have to worry about
57 restoring the full backup, then each of the incrementals in turn; an
58 incremental backup *acts* as if it's a full backup, it just takes less
61 - You can mount your bup repository as a FUSE filesystem and access the
62 content that way, and even export it over Samba.
64 - It's written in python (with some C parts to make it faster) so it's easy
65 for you to extend and maintain.
68 Reasons you might want to avoid bup
69 -----------------------------------
71 - This is a very early version. Therefore it will most probably not work
72 for you, but we don't know why. It is also missing some
73 probably-critical features.
75 - It requires python >= 2.4, a C compiler, and an installed git version >=
78 - It currently only works on Linux, MacOS X >= 10.4, or Windows (with
79 Cygwin). Patches to support other platforms are welcome.
85 - Check out the bup source code using git:
87 git clone git://github.com/apenwarr/bup
89 - Install the needed python libraries (including the development
90 libraries). On Debian or Ubuntu, this is usually:
91 apt-get install python2.6-dev python-fuse
93 Substitute python2.5-dev or python2.4-dev if you have an older system.
95 - Build the python module and symlinks:
103 (The tests should pass. If they don't pass for you, stop here and send
106 - Try making a local backup as a tar file:
108 tar -cvf - /etc | bup split -n local-etc -vv
110 - Try restoring your backup tarball:
112 bup join local-etc | tar -tf -
114 - Look at how much disk space your backup took:
118 - Make another backup (which should be mostly identical to the last one;
119 notice that you don't have to *specify* that this backup is incremental,
120 it just saves space automatically):
122 tar -cvf - /etc | bup split -n local-etc -vv
124 - Look how little extra space your second backup used on top of the first:
128 - Restore your old backup again (the ~1 is git notation for "one older than
131 bup join local-etc~1 | tar -tf -
133 - Get a list of your previous backups:
135 GIT_DIR=~/.bup git log local-etc
137 - Make a backup on a remote server (which must already have the 'bup' command
138 somewhere in the server's PATH (see /etc/profile, etc/environment,
139 ~/.profile, or ~/.bashrc), and be accessible via ssh.
140 Make sure to replace SERVERNAME with the actual hostname of your server):
142 tar -cvf - /etc | bup split -r SERVERNAME: -n local-etc -vv
144 - Try restoring the remote backup tarball:
146 bup join -r SERVERNAME: local-etc | tar -tf -
148 - Try using the new (slightly experimental) 'bup index' and 'bup save'
149 style backups, which bypass 'tar' but have some missing features (see
150 "Things that are stupid" below):
153 bup save -n local-etc /etc
155 - Do it again and see how fast an incremental backup can be:
158 bup save -n local-etc /etc
160 (You can also use the "-r SERVERNAME:" option to 'bup save', just like
161 with 'bup split' and 'bup join'. The index itself is always local,
162 so you don't need -r there.)
164 That's all there is to it!
172 bup stores its data in a git-formatted repository. Unfortunately, git
173 itself doesn't actually behave very well for bup's use case (huge numbers of
174 files, files with huge sizes, retaining file permissions/ownership are
175 important), so we mostly don't use git's *code* except for a few helper
176 programs. For example, bup has its own git packfile writer written in
179 Basically, 'bup split' reads the data on stdin (or from files specified on
180 the command line), breaks it into chunks using a rolling checksum (similar to
181 rsync), and saves those chunks into a new git packfile. There is one git
184 When deciding whether to write a particular chunk into the new packfile, bup
185 first checks all the other packfiles that exist to see if they already have that
186 chunk. If they do, the chunk is skipped.
188 git packs come in two parts: the pack itself (*.pack) and the index (*.idx).
189 The index is pretty small, and contains a list of all the objects in the
190 pack. Thus, when generating a remote backup, we don't have to have a copy
191 of the packfiles from the remote server: the local end just downloads a copy
192 of the server's *index* files, and compares objects against those when
193 generating the new pack, which it sends directly to the server.
195 The "-n" option to 'bup split' and 'bup save' is the name of the backup you
196 want to create, but it's actually implemented as a git branch. So you can
197 do cute things like checkout a particular branch using git, and receive a
198 bunch of chunk files corresponding to the file you split.
200 If you use '-b' or '-t' or '-c' instead of '-n', bup split will output a
201 list of blobs, a tree containing that list of blobs, or a commit containing
202 that tree, respectively, to stdout. You can use this to construct your own
203 scripts that do something with those values.
207 'bup index' walks through your filesystem and updates a file (whose name is,
208 by default, ~/.bup/bupindex) to contain the name, attributes, and an
209 optional git SHA1 (blob id) of each file and directory.
211 'bup save' basically just runs the equivalent of 'bup split' a whole bunch
212 of times, once per file in the index, and assembles a git tree
213 that contains all the resulting objects. Among other things, that makes
214 'git diff' much more useful (compared to splitting a tarball, which is
215 essentially a big binary blob). However, since bup splits large files into
216 smaller chunks, the resulting tree structure doesn't *exactly* correspond to
217 what git itself would have stored. Also, the tree format used by 'bup save'
218 will probably change in the future to support storing file ownership, more
219 complex file permissions, and so on.
221 If a file has previously been written by 'bup save', then its git blob/tree
222 id is stored in the index. This lets 'bup save' avoid reading that file to
223 produce future incremental backups, which means it can go *very* fast unless
224 a lot of files have changed.
227 Things that are stupid for now but which we'll fix later
228 --------------------------------------------------------
230 Help with any of these problems, or others, is very welcome. Join the
231 mailing list (see below) if you'd like to help.
233 - 'bup save' doesn't know about file metadata.
235 That means we aren't saving file attributes, mtimes, ownership, hard
236 links, MacOS resource forks, etc. Clearly this needs to be improved.
238 - There's no 'bup restore' yet.
240 'bup save' saves files in the standard git 'tree of blobs' format, so you
241 could then "restore" the files using something like 'git checkout'. But
242 that's a git command, not a bup command, so it's hard to explain and
243 doesn't support retrieving objects from a remote bup server without first
244 fetching and packing an entire (possibly huge) pack, which could be very
245 slow. Also, like 'bup save', you would need extra features in order to
246 properly restore file metadata. And files that bup has split into
247 chunks will need to be recombined. Although there's no restore tool,
248 'bup fuse' does accomplish some of this already.
250 - 'bup index' is slower than it should be.
252 It's still rather fast: it can iterate through all the filenames on my
253 600,000 file filesystem in a few seconds. But it still needs to rewrite
254 the entire index file just to add a single filename, which is pretty
255 nasty; it should just leave the new files in a second "extra index" file
258 - bup could use inotify for *really* efficient incremental backups.
260 You could even have your system doing "continuous" backups: whenever a
261 file changes, we immediately send an image of it to the server. We could
262 give the continuous-backup process a really low CPU and I/O priority so
263 you wouldn't even know it was running.
265 - bup currently has no features that prune away *old* backups.
267 Because of the way the packfile system works, backups become "entangled"
268 in weird ways and it's not actually possible to delete one pack
269 (corresponding approximately to one backup) without risking screwing up
272 git itself has lots of ways of optimizing this sort of thing, but its
273 methods aren't really applicable here; bup packfiles are just too huge.
274 We'll have to do it in a totally different way. There are lots of
275 options. For now: make sure you've got lots of disk space :)
277 - bup has never been tested on anything but Linux, MacOS, and Windows+Cygwin.
279 There's nothing that makes it *inherently* non-portable, though, so
280 that's mostly a matter of someone putting in some effort. (For a
281 "native" Windows port, the most annoying thing is the absence of ssh in
282 a default Windows installation.)
284 - bup needs better documentation.
286 According to a recent article about git in Linux Weekly News
287 (https://lwn.net/Articles/380983/), "it's a bit short on examples and
288 a user guide would be nice." Documentation is the sort of thing that
289 will never be great unless someone from outside contributes it (since
290 the developers can never remember which parts are hard to understand).
292 - bup is "relatively speedy" and has "pretty good" compression.
294 ...according to the same LWN article. Clearly neither of those is good
295 enough. We should have awe-inspiring speed and crazy-good compression.
296 Must work on that. Writing more parts in C might help with the speed.
300 Actually, that's not stupid, but you might consider it a limitation.
301 There are a bunch of Linux GUI backup programs; someday I expect someone
302 will adapt one of them to use bup.
308 bup is a work in progress and there are many ways it can still be improved.
309 If you'd like to contribute patches, ideas, or bug reports, please join the
312 You can find the mailing list archives here:
314 http://groups.google.com/group/bup-list
316 and you can subscribe by sending a message to:
318 bup-list+subscribe@googlegroups.com